sign in sign up
<<
Home , Panitumumab

Journal of Cancer Research and Clinical Oncology (2019) 145:2303–2311 https://doi.org/10.1007/s00432-019-02984-2

REVIEW – CLINICAL ONCOLOGY

A systematic review of head‑to‑head trials of approved monoclonal antibodies used in cancer: an overview of the clinical trials agenda

Jia Luo1 · Go Nishikawa1 · Vinay Prasad2,3,4

Received: 2 April 2019 / Accepted: 18 July 2019 / Published online: 8 August 2019 © Springer-Verlag GmbH Germany, part of Springer Nature 2019

Abstract Background Since 1997, several monoclonal antibodies (mAbs) targeting the same receptor or its ligand have been approved for use in oncology. However, no studies have summarized head-to-head trials of these mAbs. Methods Systematic search of the biomedical literature and ClinicalTrials.gov for randomized studies comparing mAbs targeting the same receptor or its ligand that have been completed and published, completed and unpublished, or ongoing. We extracted trial characteristics including phase, indication, enrollment or target enrollment, randomization, primary endpoint and sponsor. Results Twenty-two approved cancer mAbs had at least one other approved mAb targeting the same receptor or its ligand, totaling 41 diferent oncology indications. These include 5 anti-CD20 mAbs, 5 anti-PD1/PDL1 mAbs, 4 anti-HER2 mAbs, 3 anti-EGFR mAbs, 3 anti-VEGF mAbs and 2 anti-IL6/IL6R mAbs. Seventeen were completed and published and 14 were unpublished or ongoing trials. The completed and published trials enrolled 11,373 patients and tested 13 mAbs (13/22, 59%). Additionally, 13 (76%) contained drugs manufactured by the same company and 13 (76%) reached conclusions felt to be favorable to the sponsor. Of the 14 ongoing/completed unpublished trials, there is a total target enrollment of 3404 patients with 9 mAbs tested. Of these, 86% (12/14) are testing mAbs manufactured by the same company and 71% (10/14) are sponsored by the company that made the drug being tested. Conclusions Most trials test drugs manufactured or sponsored by the same company. An overview of clinical trials agenda may lead to more uniform testing, helping clinicians make better evidence-informed prescribing decisions.

Keywords Head-to-head trial · Monoclonal antibodies · Immunotherapy

Key points

Jia Luo and Go Nishikawa are contributed equally to this work. Question What head-to-head trials have been or are being performed for mAbs with the same target receptor or its Electronic supplementary material The online version of this article (https​://doi.org/10.1007/s0043​2-019-02984​-2) contains ligand in oncology? supplementary material, which is available to authorized users. Findings We found 17 completed and published and 14 completed and unpublished or ongoing trials of head-to-head * Vinay Prasad mAbs targeting the same receptor or its ligand. Most combi- [email protected] nations of the 22 approved mAbs and 150 possible head-to- 1 Department of Medicine, Oregon Health and Science head trials have not been tested, and most studies test drugs University, Portland, OR 97239, USA manufactured by the same company. 2 Division of Hematology Oncology, Knight Cancer Institute, Meaning For cancer mAbs with the same target, to Oregon Health and Science University, 3181 SW Sam maximize informed decision making, the overarching trials Jackson Park Road, Portland, OR 97239, USA agenda should be examined. 3 Department of Public Health and Preventive Medicine, Oregon Health and Science University, Portland, OR 97239, USA 4 Senior Scholar in the Center for Health Care Ethics, Oregon Health and Science University, Portland, OR 97239, USA

Vol.:(0123456789)1 3 2304 Journal of Cancer Research and Clinical Oncology (2019) 145:2303–2311

Introduction for each mAb, we found the most recent FDA drug label per above and identifed all randomized head-to-head trials. Since the US Food and Drug Administration’s (FDA) Next, using MEDLINE, we searched names (generic, and approval of anti-CD20 (mAb) rituxi- letter and number name e.g. MYL 1401O) of all combina- mab in 1997 for relapsed/refractory low-grade non-Hodgkin tions of pairs of mAbs and fltered by clinical trials, and a , several mAbs have been approved for use in vari- second search of the mAbs and “randomized.” We also used ous cancers. Approvals have been granted for mAbs target- Google Scholar, searching for the drug names and “phase 3” ing receptors (ErbB family EGFR and HER-2, and a second search using “randomized.” and VEGF), and recently immune modulating mAbs target- To identify all unpublished and ongoing trials comparing ing T cell activation signals (PD-1/PDL-1 and CTLA-4). mAbs head-to-head, we searched ClinicalTrials.gov (NIH The FDA has approved multiple mAbs targeting the same US National Library of Medicine 2018). For each of the que- protein that have individually shown durable clinical ben- ries, we used a combination of two mAbs and “randomized”. eft, allowing oncologists fexibility in choosing therapy for We scanned each of the trials that met our criteria which patients. included trials categorized as “recruiting,” “terminated,” Despite the approval of multiple Abs with the same tar- “active, not recruiting,” and “completed”. get, most randomized clinical trials have tested therapies We frst scanned titles, then abstracts and papers. We against placebo. Performing head-to-head randomized con- excluded trials that used experimental drugs, and those that trolled trials for mAbs would be desirable to directly com- did not test drugs head-to-head. We included trials where pare relative benefts and side efects of these interventions there were at least two arms comparing mAbs of the same (Song et al. 2009; Song et al. 2003). For this reason, we target head-to-head. A consort diagram is located in the sought to summarize all completed and upcoming trials of Supplement. head-to-head mAbs targeting the same protein. Specifcally, we aimed to describe which mAbs have been directly com- Data collection pared, how the trials were conducted, the outcomes, and a descriptive summary of the overall clinical trials agenda. For each published randomized trial, we collected names of the mAbs, trial name, phase and design, disease indication, number of patients enrolled, randomization scheme, blind- ing, treatment arms, primary endpoint, sponsor, NCT (from Methods ClinicalTrials.gov) and PMID (from MEDLINE). We classifed the conclusion of a trial as favorable to the We sought to identify all head-to-head trials comparing two sponsor if the mAb manufactured by the sponsor showed mAbs with the same target that have been performed or are the statistically superior primary endpoint. For noninferior- actively being tested in oncology. ity trials, we classifed the trial as favorable if the sponsor’s drug was non-inferior to the control arm. For trials where the sponsor tested their own drugs head-to-head, we classi- Search and inclusion criteria fed the result as favorable if the newer drug had the more favorable primary endpoint result. To identify candidate mAbs, we searched the US FDA For the ongoing head-to-head trials, we collected the fol- Drugs website (Drugs@FDA) beginning in 1997, when lowing information: names of the mAbs, phase, indication, was frst approved, for all monoclonal antibod- target enrollment, status of the trial and completion date, ies approved for a cancer indication on 1/28/18 (US Food treatment arms, primary endpoint, sponsors and collabora- and Drug Administration 2018). We collected the generic tors, and NCT. Two authors extracted the data and discrep- name, brand name, manufacturer, type of antibody (animal, ancies were discussed among all three authors to reach an human, humanized or chimeric), frst approval year, and agreement. FDA approved indication. We then included groups of mAbs that targeted the same receptor or its ligand. We searched Data analysis for and read the most recent FDA drug label at Drugs@ FDA and noted the approved indications for that drug under Descriptive statistics were used when appropriate. Tables “Indications and Usage.” were made in Microsoft Excel 2017. Figures were made in To identify all published head-to-head trials of mAbs Adobe Illustrator CC 2017. To identify the number of pos- targeting the same protein, we performed MEDLINE and sible combinations to search, we calculated the number of Google Scholar searches between 3/4/2018 and 3/22/2018. potential comparisons of pairs of mAbs with the same target We performed the search in the following four ways. First, as follows: C (n, k) = P (n, k)/k! = n!/(n–k)!, where n = total

1 3 Journal of Cancer Research and Clinical Oncology (2019) 145:2303–2311 2305 number of mAbs with the same target and k = 2. This study most commonly tested mAb was rituximab in 10 trials of head-to-head trials did not involve human subjects and did (10/17, 59%) (Fig. 1). Other common drugs included ritux- not require IRB approval. Our study was conducted between imab and hyaluronidase (5/17, 29%), (5/17, January 28, 2018 and May 7, 2018. 29%) and ado- (4/17, 24%) (Fig. 1). In total, 11,373 patients were randomized in these trials. The most common mAb patients were randomized to was Results rituximab (3174/11,373, 28%), followed by (1640/11,373, 14%) and rituximab and hyaluronidase Characteristics of included antibodies (1498/11,373, 13%). The median sample size was 500, ranging from 143 to 1418. Most of the trials were phase We found 34 mAbs approved for treating cancer from 1997 III (13/17, 76%), and the rest were two phase IB and phase to 2017. II trials (2/17, 12%) (Table 2). Eight trials tested statistical Of these, 22/34 (65%) had at least one other approved hypotheses related to drug superiority (8/17, 47%), fve mAb targeting the same receptor or its ligand (Table 1). noninferiority (5/17, 29%), two equivalence (2/17, 12%), There were fve approved anti-CD20 mAbs and anti-PD1/ one was a crossover trial (1/17, 6%) and one did not have PDL1 mAbs. We identifed four approved anti-HER2, three a formal statistical hypothesis (1/17, 6%). anti-EGFR, three anti-VEGF and two anti-IL6/IL6R mAbs. The most common indications were breast cancer (5/17, Among these, eight are human antibodies (8/22, 36%), nine 29%) and DLBCL (4/17, 23%). The most common primary humanized antibodies (9/22, 41%), four chimeric antibodies endpoint was progression free survival, used in fve trials (4/22, 18%) and one mouse antibody (1/22, 5%). (5/17, 29%). Other common primary endpoints included Defning a “unique indication” as both the cancer type ORR (4/17, 23%) and CR (either CR or pathological CR, and, if specifed, line of treatment per the FDA drug label, 4/17, 23%). Out of the 17 trials, 13 (76%) were head-to- we found each mAb is approved for between 1 and 10 indica- head trials of mAbs manufactured by the same company tions in cancer. In sum, 11 unique indications exist for the and 13 had a conclusion that favored its sponsor (13/17, anti-PD1/PDL1 mAbs, 6 for the anti-CD20 mAbs, 6 for the 76%) (Supplementary Table 3, Fig. 1). anti-HER2 mAbs, 8 for the anti-EGFR mAbs, 9 for the anti- Regarding ongoing studies, we searched a total of 174 VEGF mAbs and 1 for the anti-IL6/IL6R mAbs. studies and found 14 randomized head-to-head trials that Using the combinatorics calculation described in the met our criteria (Supplementary Table 2). The most com- methods, when we account for the number of unique indica- mon drug tested was trastuzumab in seven trials (7/14, tions for each type of mAb, the maximum number of poten- 50%), then rituximab in six trials (6/14, 43%) (Fig. 2). tial head-to-head trials is 55 for anti-PD1/PDL1 mAbs, 36 Trastuzumab and ado-trastuzumab emtansine was the most for anti-VEGF, 28 for anti-EGFR, 15 for anti-CD20, 15 for commonly tested head-to-head pair, noted in six trials anti-HER2 and 1 anti-IL6 head-to-head trials. We found 1 (6/14, 43%), followed by rituximab and obinutuzumab in anti-PD1/PDL1 [0 completed and published (C), 1 ongoing four trials (4/14, 29%). There were no head-to-head trials or completed and unpublished (O)], 1 anti-VEGF (1 C, 0 in the anti-EGFR or anti-VEGF or anti-IL6/IL6R groups. O), 1 anti-EGFR (1 C, 0 O), 19 anti-CD20 (10 C, 9 O), 13 Regarding the status of the trials, nine were recruiting anti-HER2 (5 C, 8 O), and 0 anti-IL6/IL6R trials. In short, (9/14, 64%), four terminated (4/14, 29%) and one com- out of a maximum possible number of head-to-head trials pleted (1/14, 7%). Most trials were phase II or III (12/14, of 150, we found that 31 (21%) were completed or ongoing. 86%), and the rest were phase 1 or phase 1/2 (2/14, 14%). The target enrollment ranged from 32 to 1846, with Characteristics of included trials a median of 409. Overall if all of these trials met their target, they would include 3404 patients. A network dia- We searched a total of 14,818 papers and excluded 14,801 gram of these trials shows that they are manufactured by as shown in our CONSORT diagram (Supplementary 4 companies. 12/14 (86%) of these trials test two mAbs Table 1). We found a total of 17 completed and published manufactured by the same company (Fig. 2, Supplemen- randomized head-to-head trials of a total of 13 (13/22, tary Table 3). The most common indication was breast 59%) FDA approved cancer mAbs with the same target cancer in seven trials (7/14, 50%). The most common pri- (Table 2). These trials were published between 2002 and mary endpoint was PFS (6/14, 43%), followed by DFS 2018. The most common journals were JCO (6/17, 35%) (3/14, 21%). Only one of the trials (1/14, 7%) had OS as and Lancet Haematology or Oncology (5/17, 29%). Anti- a primary endpoint. Of these studies, ten trials (10/14, CD20 mAbs had 10 head-to-head trials (10/17, 59%), anti- 71%) were sponsored by, or included as a collaborator, the HER2 mAbs had 5 trials (5/17, 29%), and EGFR mAbs company that made the mAb being tested. and anti-VEGF mAbs had 1 trial each (1/17, 6%). The

1 3 2306 Journal of Cancer Research and Clinical Oncology (2019) 145:2303–2311

Table 1 All FDA approved cancer monoclonal antibodies with the same target or its ligand Monoclonal antibody (mAb) and manufacturer Type First approval FDA approved cancer indications

Anti-CD20 mAbs Rituximab (Rituxan, Genentech/Roche) Chimeric 1997 Relapsed/refractory FL, FL with chemo, FL maintenance, DLBCL, treated CLL and untreated CLL Y90 (Zevalin, Spectrum; for- Mouse 2002 Relapsed/refractory FL, untreated FL merly IDEC) (Arzerra, Novartis) Human 2009 Untreated CLL, relapsed CLL, recurrent CLL, refractory CLL Obinutuzumab (Gazyva, Genentech/Roche) Humanized 2013 Untreated CLL, relapsed/refractory FL after rituximab, untreated FL Rituximab and hyaluronidase (Rituxan Hycela, Chimeric 2017 Relapsed/refractory FL, FL with chemo, FL maintenance, Genentech/Roche) DLBCL, treated CLL and untreated CLL Anti-HER2 mAbs Trastuzumab (Herceptin, Genentech/Roche) Humanized 1998 Adjuvant breast (± chemo), metastatic breast (± chemo), metastatic gastric (Perjeta, Roche) Humanized 2012 Untreated metastatic breast with trastuzumab, neoadju- vant breast, adjuvant breast Ado-trastuzumab emtansine (TDM1) (Kadcyla, Humanized 2013 Metastatic breast after trastuzumab Genentech/Roche) Trastuzumab-dkst (Ogivri, Mylan) Humanized 2017 Adjuvant breast (± chemo), metastatic breast (± chemo), metastatic gastric Anti-EGFR mAbs (Erbitux, Lilly) Chimeric 2004 HNSCC, metastatic HNSCC (± chemo), metastatic CRC (± chemo) Panitumumab (Vectibix, ) Human 2006 Metastatic CRC (with or after chemo) (Portrazza, Lilly) Human 2015 Squamous NSCLC Anti-VEGF mAbs (Avastin, Genentech/Roche) Humanized 2004 Metastatic CRC, NSCLC w chemo, GBM, metastatic RCC, metastatic cervical, recurrent ovarian/peritoneal/ fallopian tube (Cyramza, Lilly) Human 2014 Gastric (± chemo), metastatic NSCLC after chemo, metastatic CRC​ Bevacizumab-awwb (Mvasi, Amgen) Humanized 2017 Metastatic CRC (frst line or after bevacizumab progres- sion), NSCLC with chemo, GBM, metastatic RCC, metastatic cervical Anti-IL6/IL6R mAbs Tocilizumab (Actemra, Genentech/Roche) Humanized 2010 Cytokine release syndrome (Sylvant, Janssen) Chimeric 2014 Multicentric Castleman disease Anti-PD1/PDL1 mAbs (Opdivo, BMS) Human 2014 Metastatic (± ), adjuvant mela- noma, metastatic NSCLC, advanced RCC, HL, meta- static HNSCC, metastatic bladder, MSI-H or dMMR CRC, HCC (Keytruda, Merck) Human 2015 Metastatic melanoma, metastatic NSCLC, metastatic HNSCC, refractory HL (Tecentriq, Genentech/Roche) Humanized 2016 Metastatic bladder, metastatic NSCLC (Bavencio, Pfzer) Human 2017 Metastatic merkel cell carcinoma (Imfnzi, AstraZeneca) Human 2017 Metastatic bladder

NSCLC non small cell lung cancer, RCC ​renal cell, HL hodgkin lymphoma, HNSCC head and neck squamous cell cancer, MSI-H microsatellite instability-high, dMMR mismatch repair defcient, CRC​ , HCC hepatocellular carcinoma, FL follicular lymphoma, DLBCL dif- fuse large B-cell lymphoma, CLL chronic lymphocytic , GBM glioblastoma multiforme

1 3 Journal of Cancer Research and Clinical Oncology (2019) 145:2303–2311 2307 (Witzig et al. et al. (Witzig 2002 ) (Goede et al. (Goede et al. 2014 ) (Vitolo et al. et al. (Vitolo 2017 ) (Marcus et al. et al. (Marcus 2017 ) Imhof et al. Imhof et al. 2017 ) (Salar et al. (Salar et al. 2014 ) (Assouline 2016 ) et al. (Davies et al. et al. (Davies 2017 ) (Lugtenburg (Lugtenburg 2017 ) et al. (Rummel (Rummel 2017 ) et al. PMID 12011122 24401022 28796588 28976863 28029326 (van 28029326 (van 24821885 26947201 28476440 28935843 28031173 NCT None NCT01010061 NCT01287741 NCT01332968 NCT01014208 NCT00930514 NCT01292603 NCT01200758 NCT01649856 NCT01724021 Conclusion favoring sponsor Yes Yes No Yes No Yes Yes Yes Yes Yes Roche Roche Roche Genmab, Novartis Roche Roche Roche Roche Roche Sponsor IDEC* Genentech/ Genentech/ Genentech/ GSK, Genentech/ Genentech/ Genentech/ Genentech/ Genentech/ trough trough - concentra tion centration centration 5 cycle of induc - tion preference preference survey Primary endpoint ORR PFS PFS PFS PFS Serum Serum con - ORR at end CR Patient Patient sQ R)- > G-C vs R–C G-C vs CHOP (G-CHOP) rituximab-vs CHOP (R-CHOP) R-CHOP/R-CVP vs R-DHAP vs R-fu-Cy sQ R-CVP vs vs sQ R-CVP R-CHOP/R-CVP R-CHOP CHOP/CVP/B CHOP/CVP/B (R - vs Arms Y90 IT vs R Y90 IT vs Chlorambucil (C) vs (C) vs Chlorambucil Obinituzumab- G-CHOP/G-CVP vs G-CHOP/G-CVP vs Ofatumumab-DHAP Ofatumumab-DHAP sQ R vs R sQ R vs sQ R-fu-Cy vs vs sQ R-fu-Cy sQ R-CHOP/ sQ R-CHOP vs vs sQ R-CHOP (R - > sQ R - > R)- Blinding Open label Open label Open label Open label Open label Open label Open label Open label Open label Open label - n and rand omization 143, 1:1 787, 1:2:2 1418, 1:1 1202, 1:1 447, 1:1 154, 1:1 176, 1:1 410, 1:1 572, 2:1 743, 1:1 low grade FL or low NHL transformed stage DLBCL stage NHL DLBCL Relapsed/refractory Indication Untreated CLL Untreated Untreated advanced advanced Untreated Untreated indolent Untreated Relapsed/refractory Maintenance FL Untreated CLL Untreated Untreated FL Untreated Untreated DLBCL Untreated Untreated DLBCL Untreated - noninferi ority - noninferi ority - noninferi ority “no formal “no formal statistical - hypoth esis” crossover Phase III Phase, Design Phase III Phase III Phase III Phase III Phase IB, Phase IB, Phase III, Phase III, Phase III, 20:2453–63 2014;370:1101–10 35:3529–37 2017;377:1331–44 2016 35:544–551 32:1782–91 Haem 2016 3:e128–38 Haem 2017 4:e272–282 matologica 2017 matologica 102:1913–1922 2017 28:836–842 JCO 2002 JCO Trial CLL11; NEJM GOYA; JCO 2017 JCO GOYA; GALLIUM; NEJM GALLIUM; ORCHARRD; JCO JCO ORCHARRD; JCO 2014 JCO SAWYER; Lancet Lancet SAWYER; SABRINA; Lancet Lancet SABRINA; MabEASE; Hae - PrefMab; Ann Onc PrefMab; (Rituxan, (Rituxan, Genentech/ Roche) Rituximab mAb 2 mAb Rituximab Rituximab Rituximab Rituximab Rituximab Rituximab Rituximab Rituximab Rituximab Summary of published head-to-head trials of FDA approved cancer monoclonal antibodies approved Summary trials of published head-to-head of FDA momab tiuxetan (Y90 IT) (Zevalin, Spectrum; formerly IDEC) (Gazyva, (Gazyva, Genentech/ Roche) zumab zumab (Arzerra, Novartis) - hyaluroni dase (sQ R) (Rituxan Hycela, Genentech/ Roche) - hyaluroni dase - hyaluroni dase - hyaluroni dase - hyaluroni dase Y90 ibritu - Obinutuzumab Obinutuzumab Obinutu - Obinutu - Ofatumumab Ofatumumab Rituximab and Rituximab Rituximab and Rituximab Rituximab and Rituximab Rituximab and Rituximab Rituximab and Rituximab 2 Table mAb 1 mAb Anti-CD 20 mAbs

1 3 2308 Journal of Cancer Research and Clinical Oncology (2019) 145:2303–2311 - (Gianni et al. (Gianni et al. 2012 ) (Perez et al. et al. (Perez 2017 ) beck et al. et al. beck 2017 ) (Hurvitz et al. 2018 ) (Price et al. 2014 ) (Rugo et al. et al. (Rugo 2017 ) PMID 22153890 28056202 28682681 (Har None 29175149 24739896 27918780 NCT NCT00545688 NCT01120184 NCT01817452 NCT01966003 NCT02131064 NCT01001377 NCT02472964 Conclusion favoring sponsor No Yes Yes Yes No Yes Yes Roche Roche Roche Roche Sponsor Genentech/ Genentech/ Genentech/ Genentech/ Amgen Amgen Mylan week 24 week Primary endpoint PathCR PFS PathCR PathCR ORR OS ORR at etaxel vs vs etaxel pertuzumab- + tras tuzumab + doc - vs etaxel pertuzumab- + tras pertutuzumab vs - zumab + docetaxel ane vs TDM1 vs TDM1 vs ane vs - TDM1 + trastu zumab - trastu vs therapy zumab + endocrine therapy vs docetaxel, docetaxel, vs carboplatin, trastuzumab, pertuzumab vs bevacizumab, bevacizumab, vs carbo-taxel cetuximab taxane + taxane dkst vs + tax trastuzumab - ane Arms + doc - Trastuzumab + tax - Trastuzumab TDM1 ± endocrine TDM1 pertuzumab ABP 215, carbo-taxel ABP 215, carbo-taxel Panitumumab vs vs Panitumumab Trastuzumab- Blinding Open label Open label Open label Open label Open label Open label Open label - n and rand omization 417, 1:1:1:1 1095, 1:1:1 375, 1:1 444, 1:1 642, 1:1 1010, 1:1 500, 1:1 naïve breast cancer breast naïve breast cancer breast breast cancer breast breast cancer breast NSCLC KRAS metastatic colorectal cancer treatment HER2 + treatment Indication advanced HER2 + advanced early HER2 + HR early stage HER2 + stage II–III Non-squamous Non-squamous Chemo refractory WT Metastatic breast - noninferi ority - equiva lence - noninferi ority - equiva lence Phase II Phase, Design Phase III, Phase II Phase III Phase III, Phase III, Phase III, Onc 2012 13:25–32 2017 35:141–148 35:3046–54 Lancet Onc 2018 Lancet 19:115–26 (abstract only) (abstract Oncol 2014; 15:569–79 2017 317:37–47 NeoSphere; Lancet Lancet NeoSphere; Trial MARIANNE; JCO MARIANNE; JCO ADAPT; JCO 2017 JCO ADAPT; KRISTINE; KRISTINE; JTO 2017 P2.03a-025 JTO ASPECCT; Lancet Lancet ASPECCT; HERiTAge; JAMA JAMA HERiTAge; (Herceptin, (Herceptin, Genentech/ Roche) (Avastin, (Avastin, Genentech/ Roche) (Erbitux, Lilly) Trastuzumab Trastuzumab mAb 2 mAb Trastuzumab Trastuzumab Trastuzumab Bevacizumab Bevacizumab Cetuximab Cetuximab Trastuzumab (continued) (Perjeta, (Perjeta, Genentech/ Roche) tuzumab emtansine (TDM1) (Kadcyla, Genentech/ Roche) tuzumab emtansine tuzumab emtansine awwb awwb (ABP 215) (Mvasi, Amgen) (Vectibix, (Vectibix, Amgen) dkst dkst (Ogviri, Mylan) Pertuzumab - Ado-tras - Ado-tras - Ado-tras Bevacizumab- Panitumumab Panitumumab Trastusumab- 2 Table IDEC pharmaceuticals the by publication of this*At trial, manufactured and rituximab both Y90 ibritumomab were tiuxetan mAb 1 mAb Anti-HER2 mAbs Anti-VEGF mAbs Anti-VEGF Anti-EGFR mAbs

1 3 Journal of Cancer Research and Clinical Oncology (2019) 145:2303–2311 2309

etan (Spectrum, formerly IDEC)

trastuzumab-dkst rituximab and (Mylan) hyaluronidase (Genentech/ Roche) s

ofatumumab (Nova ado-trastuzumab emtansine (TDM1) trastuzumab (Genentech/ Roche) (Genentech/ Roche)

rituximab obinutuzumab (Genentech/ Roche and (Genentech/ Roche) pertuzumab IDEC) (Genentech/ Roche)

cetuximab Spectrum, (Lilly) formerly IDEC

Genentech/ Roche

EGFR mAbs n=5000 GF mAbs Amgen panitumumab (Amgen) Lilly bevacizumab n=500 (Genentech/ Roche) bevacizumab-awwb n=50 (Amgen) Mylan not favorable to sponsor Nova s favorable to sponsor

Fig. 1 Network diagram of completed and published head to head tri- coded by manufacturer. Each line connecting two circles represents als of mAbs of the same target or its ligand. We identifed 17 of these one head to head trial, and the color designates whether the result trials. The area of each node represents the approximate total number favors the sponsor, as defned in the methods of patients who received the drug in these studies. Nodes are color

Discussion comparisons in other felds of medicine (Flacco et al. 2015; Estellat and Ravaud 2012; Buesching et al. 2012). While we There are a limited number of head‑to‑head trials do not believe that all potential head-to-head trials ought to comparing approved oncology mAbs be performed, it is notable that a minority of comparisons have been or are being attempted. We found few head-to-head trials of monoclonal antibodies Out of a total of 150 combinations of head-to-head tri- that are FDA approved in oncology. In the published litera- als that could be performed due to unique approved FDA ture, we found 17 completed and published head-to-head tri- indications, we found that only 21% (31/150) have been or als of these drugs, though more combinations were possible are being undertaken. While not all of these trials should be (Table 2). All of the trials were industry sponsored by one performed given the cost of testing every possible combina- of the mAb manufacturers, and the majority of these trials tion, the published literature is predominantly represented by were comparing two mAbs manufactured by the same spon- mAbs made by the sponsor, with results favoring the spon- sor (13/17, 76%) (Fig. 1, Supplementary Table 3). A search sor. Of the 17 completed trials, the majority (13/17, 76%) through trials in the ClinicalTrials.gov database of recently reach results that favor the sponsor (Fig. 1, Table 1). completed and ongoing trials yielded 14 randomized head- to-head trials and 12/14 (86%) of these compared mAbs Inefciencies may exist in the research agenda manufactured by the same sponsor (Supplementary Table 3, testing mAbs Fig. 1). In sum, these 21 studies in total test 15 of the 22 (68%) mAbs in cancer that have multiple approvals with These results suggest that despite the fact that there are many the same target. Interestingly there are no published and randomized trials of mAbs in oncology, there may be inef- only 1 ongoing head-to-head trial of approved anti-PD1/ fciencies in selecting the clinical trials being performed PD-L1 mAbs, despite the popularity of these drugs. This (Ioannidis 2016), similar to what has been previously shown fnding is similar to studies that show a lack of head-to-head in the literature for diferent anti-cancer drugs (Carlisle et al.

1 3 2310 Journal of Cancer Research and Clinical Oncology (2019) 145:2303–2311

ofatumumab ado-trastuzumab (Nova rituximab and hyaluronidase trastuzumab emtansine (Genentech/ Roche) (Genentech/ Roche) (Genentech/ Roche)

s

pertuzumab (Genentech/ Roche) obinutuzumab (Genentech/ Roche)

rituximab Merck (Genentech/ Roche) Genentech/ Roche

n=5000 nivolumab Nova s (BMS)

pembrolizumab BMS (Merck) n=500 n=50

head to head trial

Fig. 2 Network diagram of ongoing and completed but not pub- in these studies (target enrollment divided by the number of arms). lished head to head trials of mAbs of the same target or its ligand. Nodes are color coded by manufacturer. Each line connecting two We identifed 14 of these trials. The area of each node represents the nodes represents one head to head trial approximate number of patients who received or will receive the drug

2016; Mattina et al. 2017). Our review indicates that there accordance with standardized systematic review reporting are additional inefciencies when looking at the sum total criteria. of head-to-head trials in cancer. Second, some of the FDA approved drugs, especially those Per the National Academy of Medicine’s (formerly Insti- in the PD1/PDL1 family, were only recently approved in oncol- tute of Medicine) Initial National Priorities for Compara- ogy–thus, limiting the amount of time during which compari- tive Efectiveness Research, “Once an intervention has been sons could be attempted. However, a recent study found over shown to be efective against a placebo, head-to-head trials 1000 ongoing trials of anti-PD1/PDL1 mAbs with other agents address the critical question ‘What works best for whom?’” with target enrollment of thousands of patients (Tang et al. (Institute of Medicine 2009). In our study, we found that in 2018), suggesting that even though early in the lifecycle of the mAb space, important head-to-head trials may not be this class of mAbs, there is potential for ongoing comparisons. conducted. This is particularly interesting when contrasting Third, our study only examined mAbs targeting the same this agenda against a recent study that found over 1000 trials receptor or its ligand. We are missing head-to-head com- testing the new immune modulating mAbs in diferent com- parisons between mAbs and other targeted agents. However, binations. Thus, although there are many trials of checkpoint even when focused on mAbs themselves, we have identifed inhibitors, few seek to ascertain if one antibody is superior notable patterns. Future research may consider examination to the other, or ofer diferent side efect profles. Given that of broader collections of agents. antibodies are biological molecules, prone to idiosyncratic diferences, this omission is notable. Conclusion Limitations Our study found that mAbs used in cancer targeting the same There are several limitations to our work. First, it is possible receptor or ligand are often not tested in head-to-head rand- we did not identify all head-to-head trials of mAbs targeting omized trials; and if they were completed and published, 76% the same receptor or its ligand. However, we systematically (13/17) reach conclusions that favored the sponsoring com- performed multiple MEDLINE and Google Scholar searches pany. Consideration of the overarching clinical trials agenda and looked on ClinicalTrials.gov for additional studies in of mAbs may lead to more rational portfolios.

1 3 Journal of Cancer Research and Clinical Oncology (2019) 145:2303–2311 2311

Compliance with ethical standards Marcus R, Davies A, Ando K et al (2017) Obinutuzumab for the frst-line treatment of follicular lymphoma. N Engl J Med 377(14):1331–1344 Mattina J, Carlisle B, Hachem Y, Fergusson D, Kimmelman J (2017) Conflict of interest Dr Prasad reports receiving royalties from his book Inefciencies and patient burdens in the development of the tar- Ending Medical Reversal; that his work is funded by the Laura and geted cancer drug : a systematic review. PLoS Biol John Arnold Foundation; that he has received honoraria for Grand 15(2):e2000487 Rounds/lectures from several universities, medical centers, and profes- NIH US National Library of Medicine (2018) ClinicalTrials.gov. https://​ sional societies and payments for contributions to Medscape. Drs. Luo clini​caltr​ials.gov/. Accessed 1 June 2018 and Nishikawa have no confict of interest. Perez EA, Barrios C, Eiermann W et al (2017) Trastuzumab emtansine with or without pertuzumab versus trastuzumab plus taxane for References human epidermal 2-positive, advanced breast cancer: primary results from the phase III MARIANNE study. J Clin Assouline S, Buccheri V, Delmer A et al (2016) Pharmacokinetics, Oncol 35(2):141–148 safety, and efcacy of subcutaneous versus intravenous rituximab Price TJ, Peeters M, Kim TW et al (2014) Panitumumab versus cetuxi- plus as treatment for chronic lymphocytic leukaemia mab in patients with chemotherapy-refractory wild-type KRAS exon (SAWYER): a phase 1b, open-label, randomised controlled non- 2 metastatic colorectal cancer (ASPECCT): a randomised, multi- inferiority trial. Lancet Haematol 3(3):e128–e138 centre, open-label, non-inferiority phase 3 study. Lancet Oncol. Buesching DP, Luce BR, Berger ML (2012) The role of private indus- 15(6):569–579 try in pragmatic comparative efectiveness trials. J Comp Ef Res Rugo HS, Barve A, Waller CF et al (2017) Efect of a proposed trastu- 1(2):147–156 zumab biosimilar compared with trastuzumab on overall response Carlisle B, Demko N, Freeman G et al (2016) Beneft, risk, and outcomes rate in patients with ERBB2 (HER2)-positive metastatic breast can- in drug development: a systematic review of . J Natl Cancer cer: a randomized clinical trial. JAMA 317(1):37–47 Inst 108(1):djv292 Rummel M, Kim TM, Aversa F et al (2017) Preference for subcutane- Davies A, Merli F, Mihaljevic B et al (2017) Efcacy and safety of sub- ous or intravenous administration of rituximab among patients with cutaneous rituximab versus intravenous rituximab for frst-line untreated CD20+ difuse large B-cell lymphoma or follicular lym- treatment of follicular lymphoma (SABRINA): a randomised, open- phoma: results from a prospective, randomized, open-label, crosso- label, phase 3 trial. Lancet Haematol 4(6):e272–e282 ver study (PrefMab). Ann Oncol 28(4):836–842 Estellat C, Ravaud P (2012) Lack of head-to-head trials and fair control Salar A, Avivi I, Bittner B et al (2014) Comparison of subcutaneous arms: randomized controlled trials of biologic treatment for rheu- versus intravenous administration of rituximab as maintenance treat- matoid arthritis. Arch Intern Med 172(3):237–244 ment for follicular lymphoma: results from a two-stage, phase IB Flacco ME, Manzoli L, Boccia S et al (2015) Head-to-head randomized study. J Clin Oncol 32(17):1782–1791 trials are mostly industry sponsored and almost always favor the Song F, Altman DG, Glenny AM, Deeks JJ (2003) Validity of indi- industry sponsor. J Clin Epidemiol 68(7):811–820 rect comparison for estimating efcacy of competing interven- Gianni L, Pienkowski T, Im YH et al (2012) Efcacy and safety of tions: empirical evidence from published meta-analyses. BMJ neoadjuvant pertuzumab and trastuzumab in women with locally 326(7387):472 advanced, infammatory, or early HER2-positive breast cancer (Neo- Song F, Loke YK, Walsh T, Glenny AM, Eastwood AJ, Altman DG Sphere): a randomised multicentre, open-label, phase 2 trial. Lancet (2009) Methodological problems in the use of indirect comparisons Oncol 13(1):25–32 for evaluating healthcare interventions: survey of published system- Goede V, Fischer K, Busch R et al (2014) Obinutuzumab plus chloram- atic reviews. BMJ 338:b1147 bucil in patients with CLL and coexisting conditions. N Engl J Med Tang J, Shalabi A, Hubbard-Lucey VM (2018) Comprehensive analysis 370(12):1101–1110 of the clinical immuno-oncology landscape. Ann Oncol 29(1):84–91 Harbeck N, Gluz O, Christgen M et al (2017) De-escalation strategies in US Food and Drug Administration (2018) Drugs@FDA: FDA approved human receptor 2 (HER2)-positive early drug products. https​://www.acces​sdata​.fda.gov/scrip​ts/cder/daf/. breast cancer (BC): fnal analysis of the west german study group Accessed 1 June 2018 adjuvant dynamic marker-adjusted personalized therapy trial opti- van Imhof GW, McMillan A, Matasar MJ et al (2017) Ofatumumab ver- mizing risk assessment and therapy response prediction in early BC sus rituximab salvage chemoimmunotherapy in relapsed or refrac- HER2- and hormone receptor-positive phase II randomized trial- tory difuse large B-cell lymphoma: the ORCHARRD study. J Clin efcacy, safety, and predictive markers for 12 weeks of neoadjuvant Oncol 35(5):544–551 trastuzumab emtansine with or without endocrine therapy (ET) ver- Vitolo U, Trneny M, Belada D et al (2017) Obinutuzumab or rituximab sus trastuzumab plus ET. J Clin Oncol 35(26):3046–3054 plus cyclophosphamide, doxorubicin, vincristine, and prednisone in Hurvitz SA, Martin M, Symmans WF et al (2018) Neoadjuvant trastu- previously untreated difuse large B-cell lymphoma. J Clin Oncol zumab, pertuzumab, and chemotherapy versus trastuzumab emtan- 35(31):3529–3537 sine plus pertuzumab in patients with HER2-positive breast cancer Witzig TE, Gordon LI, Cabanillas F et al (2002) Randomized controlled (KRISTINE): a randomised, open-label, multicentre, phase 3 trial. trial of yttrium-90-labeled ibritumomab tiuxetan radioimmunother- Lancet Oncol 19(1):115–126 apy versus rituximab immunotherapy for patients with relapsed or Institute of Medicine (2009) Initial national priorities for comparative refractory low-grade, follicular, or transformed B-cell non-Hodg- efectiveness research. The National Academies Press, Washington, kin’s lymphoma. J Clin Oncol 20(10):2453–2463 DC Ioannidis JPA (2016) Why most clinical research is not useful. PLoS Publisher’s Note Springer Nature remains neutral with regard to Med 13(6):e1002049 jurisdictional claims in published maps and institutional afliations. Lugtenburg P, Avivi I, Berenschot H et al (2017) Efcacy and safety of subcutaneous and intravenous rituximab plus cyclophosphamide, doxorubicin, vincristine, and prednisone in frst-line difuse large B-cell lymphoma: the randomized MabEase study. Haematologica 102(11):1913–1922

1 3